Healthcare systems have progressively shifted from paternalistic models toward more inclusive, participatory paradigms that emphasize patient-, person-, and people-centered care. These approaches position individuals as active agents in decisions affecting their health and foreground two key concepts: shared decision-making (SDM) and individual preferences. Within this evolving landscape, this thesis advances the empirical and methodological understanding of shared-decision making and preference elicitation through three complementary essays. The first essay investigates how and why digital patient decision aids (PtDAs) work in cancer treatment SDM using a rapid realist review. This approach focuses on identifying context–mechanism–outcome (CMO) configurations to explain how and under what conditions digital PtDAs influence decision processes and outcomes. Forty-two articles covering 30 distinct tools were analyzed, with evidence translated into “if–then” statements to uncover the potential links between any two of the contexts, mechanisms, and outcomes. The synthesis yielded ten CMO configurations, suggesting that factors such as patients’ individual characteristics (e.g., emotional status, learning needs, prior experience with cancer), the presence of a support network, patients’ perceptions of the appropriateness of the PtDAs, and healthcare professionals’ behavior significantly shape patient engagement with these tools, thereby influencing both the SDM process and its quality. The findings underscore that digital PtDAs are not universally effective, highlighting the need for tailored interventions. The second essay presents a comprehensive protocol for evaluating an artificial intelligence (AI)-driven intervention, the CINDERELLA APProach, designed to support SDM among breast cancer patients undergoing locoregional treatments. The intervention uses AI to generate personalized visual simulations of post-surgical aesthetic outcomes via a mobile application, aiming to align expectations with actual results and improve psychosocial wellbeing and quality of life. Given the complexity and multidimensional nature of the intervention, the proposed evaluation framework extends beyond conventional assessments of clinical effectiveness by systematically integrating economic, financial, implementation, and environmental dimensions. Overall, this essay proposes a comprehensive and methodologically rigorous framework for evaluating complex digital health interventions, addressing key gaps in the literature on PtDAs and AI applications in healthcare. The third essay investigates preferences for seasonal influenza vaccination among the Italian general population through a discrete choice experiment (DCE), conducted on a representative sample of 1,203 adults. The study quantifies preferences for key vaccine attributes (i.e., characteristics), estimates willingness-to-pay, and predicts vaccination uptake. Results from mixed logit model indicate that all included attributes significantly influence decision between vaccination and non-vaccination, with out-of-pocket cost and the risk of serious adverse events emerging as the most influential factors. Substantial preference heterogeneity is observed, driven by several socio-demographic characteristics, prior vaccination behavior, and individual beliefs. Predicted uptake varies considerably, reaching approximately 70% under optimal attribute levels for an average individual. Overall, the findings suggest that heterogeneity in vaccination preferences require targeted communication strategies to address safety misconceptions about vaccines and policies aimed at reducing financial barriers, such as expanding free vaccination, to effectively increase uptake.
I sistemi sanitari hanno progressivamente abbandonato modelli paternalistici a favore di paradigmi più inclusivi e partecipativi, che pongono al centro il paziente, la persona e la collettività. Tali approcci considerano gli individui come attori attivi nelle decisioni riguardanti la propria salute e valorizzano due concetti chiave: il processo decisionale condiviso (shared decision-making, SDM) e le preferenze individuali. In questo contesto in evoluzione, la presente tesi contribuisce ad approfondire la comprensione empirica e metodologica dello SDM e dell’elicitazione delle preferenze attraverso tre essay complementari. Il primo essay analizza come e perché gli strumenti digitali a supporto delle decisioni del paziente (patient decision aids, PtDAs) funzionano nel contesto dello SDM in oncologia, mediante una rapid realist review. Questo approccio mira a identificare configurazioni contesto–meccanismo–esito (CMO) per spiegare come e in quali condizioni i PtDAs digitali influenzano i processi decisionali e i relativi esiti. Sono stati analizzati 42 articoli relativi a 30 strumenti distinti, traducendo le evidenze in affermazioni “se–allora” per individuare i possibili legami tra contesti, meccanismi ed esiti. La sintesi ha prodotto dieci configurazioni CMO, evidenziando come fattori quali le caratteristiche individuali dei pazienti (ad esempio stato emotivo, bisogni informativi, esperienza pregressa con il cancro), la presenza di una rete di supporto, la percezione di adeguatezza dei PtDAs e il comportamento dei professionisti sanitari influenzino significativamente il coinvolgimento dei pazienti e, di conseguenza, sia il processo di SDM sia la sua qualità. I risultati sottolineano che i PtDAs digitali non sono universalmente efficaci, evidenziando la necessità di interventi personalizzati. Il secondo essay presenta un protocollo completo per la valutazione di un intervento basato sull’intelligenza artificiale (IA), il CINDERELLA APProach, progettato per supportare lo SDM tra le pazienti affette da tumore al seno sottoposte a trattamenti locoregionali. L’intervento utilizza l’IA per generare, tramite un’applicazione mobile, simulazioni visive personalizzate degli esiti estetici post-chirurgici, con l’obiettivo di allineare le aspettative ai risultati effettivi e migliorare il benessere psicosociale e la qualità della vita. Considerata la complessità e la natura multidimensionale dell’intervento, il framework di valutazione proposto va oltre le tradizionali analisi di efficacia clinica, integrando sistematicamente dimensioni economiche, finanziarie, implementative e ambientali. Nel complesso, questo essay propone un approccio metodologicamente rigoroso e completo per la valutazione di interventi digitali complessi in sanità, colmando importanti lacune nella letteratura sui PtDAs e sulle applicazioni dell’IA in ambito sanitario. Il terzo essay esamina le preferenze per la vaccinazione antinfluenzale stagionale nella popolazione generale italiana attraverso un esperimento a scelta discreta (discrete choice experiment, DCE), condotto su un campione rappresentativo di 1.203 adulti. Lo studio quantifica le preferenze per le principali caratteristiche dei vaccini, stima la disponibilità a pagare e l’adesione prevista alla vaccinazione. I risultati del modello mixed logit indicano che tutti gli attributi considerati influenzano significativamente la scelta tra vaccinazione e non vaccinazione, con il costo out-of-pocket e il rischio di eventi avversi gravi che emergono come i fattori più rilevanti. Si osserva un’elevata eterogeneità delle preferenze, influenzata da caratteristiche socio-demografiche, precedenti comportamenti vaccinali e credenze individuali. L’adesione prevista varia considerevolmente, raggiungendo circa il 70% in presenza di livelli ottimali degli attributi del vaccino per un individuo medio. Nel complesso, i risultati suggeriscono che l’eterogeneità delle preferenze richiede strategie di comunicazione mirate per contrastare le percezioni errate sulla sicurezza dei vaccini e politiche volte a ridurre le barriere economiche, come l’estensione della gratuità della vaccinazione, al fine di aumentare efficacemente la copertura vaccinale.
Borsoi, Ludovica, METHODS TO INVESTIGATE SHARED DECISION-MAKING AND INDIVIDUAL PREFERENCES IN HEALTHCARE: THREE ESSAYS, Ciani, Oriana, Università Cattolica del Sacro Cuore MILANO:Ciclo XXXVII [https://hdl.handle.net/10807/332976]
METHODS TO INVESTIGATE SHARED DECISION-MAKING AND INDIVIDUAL PREFERENCES IN HEALTHCARE: THREE ESSAYS
Borsoi, Ludovica
2026
Abstract
Healthcare systems have progressively shifted from paternalistic models toward more inclusive, participatory paradigms that emphasize patient-, person-, and people-centered care. These approaches position individuals as active agents in decisions affecting their health and foreground two key concepts: shared decision-making (SDM) and individual preferences. Within this evolving landscape, this thesis advances the empirical and methodological understanding of shared-decision making and preference elicitation through three complementary essays. The first essay investigates how and why digital patient decision aids (PtDAs) work in cancer treatment SDM using a rapid realist review. This approach focuses on identifying context–mechanism–outcome (CMO) configurations to explain how and under what conditions digital PtDAs influence decision processes and outcomes. Forty-two articles covering 30 distinct tools were analyzed, with evidence translated into “if–then” statements to uncover the potential links between any two of the contexts, mechanisms, and outcomes. The synthesis yielded ten CMO configurations, suggesting that factors such as patients’ individual characteristics (e.g., emotional status, learning needs, prior experience with cancer), the presence of a support network, patients’ perceptions of the appropriateness of the PtDAs, and healthcare professionals’ behavior significantly shape patient engagement with these tools, thereby influencing both the SDM process and its quality. The findings underscore that digital PtDAs are not universally effective, highlighting the need for tailored interventions. The second essay presents a comprehensive protocol for evaluating an artificial intelligence (AI)-driven intervention, the CINDERELLA APProach, designed to support SDM among breast cancer patients undergoing locoregional treatments. The intervention uses AI to generate personalized visual simulations of post-surgical aesthetic outcomes via a mobile application, aiming to align expectations with actual results and improve psychosocial wellbeing and quality of life. Given the complexity and multidimensional nature of the intervention, the proposed evaluation framework extends beyond conventional assessments of clinical effectiveness by systematically integrating economic, financial, implementation, and environmental dimensions. Overall, this essay proposes a comprehensive and methodologically rigorous framework for evaluating complex digital health interventions, addressing key gaps in the literature on PtDAs and AI applications in healthcare. The third essay investigates preferences for seasonal influenza vaccination among the Italian general population through a discrete choice experiment (DCE), conducted on a representative sample of 1,203 adults. The study quantifies preferences for key vaccine attributes (i.e., characteristics), estimates willingness-to-pay, and predicts vaccination uptake. Results from mixed logit model indicate that all included attributes significantly influence decision between vaccination and non-vaccination, with out-of-pocket cost and the risk of serious adverse events emerging as the most influential factors. Substantial preference heterogeneity is observed, driven by several socio-demographic characteristics, prior vaccination behavior, and individual beliefs. Predicted uptake varies considerably, reaching approximately 70% under optimal attribute levels for an average individual. Overall, the findings suggest that heterogeneity in vaccination preferences require targeted communication strategies to address safety misconceptions about vaccines and policies aimed at reducing financial barriers, such as expanding free vaccination, to effectively increase uptake.| File | Dimensione | Formato | |
|---|---|---|---|
|
tesi_completa_Borsoi.pdf
non disponibili
Licenza:
Non specificato
Dimensione
1.78 MB
Formato
Adobe PDF
|
1.78 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.



